The UK civil service isn’t ready for AI yet – but there are reasons for optimism
Global Government Forum’s Digital Leaders Study 2024, conducted with knowledge partner Cognizant, gathered views from top Whitehall leaders about the state of AI and data in government today.
At a time when AI is never out of the headlines, Global Government Forum’s study examined how ready government was to implement it.
I was a co-author of the report with Kevin Cunnington and Yatin Mahandru. Kevin is a senior UK digital leader who spent his early career in programming and IT consultancy, later becoming the global head of online for Vodafone Group and director general of the Business Transformation Group at the Department for Work and Pensions (DWP). He was then also director general of the UK’s Government Digital Service (GDS) from August 2016 to July 2019, and digital envoy for the UK and director general of the International Government Service until 2021.Yatin Mahandru is vice president and head of public sector and health, UK and Ireland for Cognizant, one of the largest technology services companies in the world, and Cognizant was GGF’s knowledge partner for this project
Here’s a snapshot of what we learned from those at the top of the civil service about this important topic.
Whitehall is driving some of the AI agenda – but implementation is limited
Speaking to digital leaders across government – from the Treasury and Home Office, to NHS England and the Government Digital Service (GDS) – offered us insights into the varied AI picture in the system today. It was clear that there’s excitement about the possibilities of this technology and that has been further encouraged by leaders’ own use of freely available generative AI tools.
We heard that Whitehall has been genuinely world-leading on AI safety. The AI Safety Summit and AI Safety Institute are both strong examples of this. The UK has been setting standards and convening the right stakeholders for what is an essential conversation about using these tools safely.
But one interviewee’s comment – which offered the defining metaphor of our report – really cut through. They told us that, when it comes to implementing AI within government: “It’s like we’re driving a car with one foot very firmly pressed down on the accelerator and one foot pressed down on the brake.”
What they meant by this was that the UK is a leader on standards and safety (‘the brake’) and, at the same time, there is considerable hype and excitement driving some limited activity in government (‘the accelerator’).
The combination of the two means that things aren’t moving fast enough in terms of practical implementation. In reality, Whitehall has gone much further on safety than on scaling up AI so far. That contrasts with some other governments, like Singapore, which have begun rolling out AI applications much more quickly.
Download the Digital Leaders Study 2024 report
AI maturity is low – with no central vision yet available
This became even more evident when we used the Seven Lenses of Transformation to ask digital leaders about AI maturity. We heard a relatively pessimistic response.
Vision – whether there is a clear, compelling, cross-government understanding of what AI should be used for – is the fundamental first component of the Seven Lenses framework. Without a core vision, the ability to implement AI successfully or build the right workforce mix to enable it just isn’t possible.
And in Whitehall, we heard that this vision is indeed lacking. That’s partly because AI leadership within government was previously too fragmented – until recent changes which introduced a single AI voice. It’s also because an AI vision which had been in development through the Central Digital and Data Office (CDDO) hasn’t yet received political sign-off for publication.
Whatever the reasons, without this clear vision, the even more challenging work of actually implementing AI has proven limited so far.
The fundamentals – especially data sharing – need to be strengthened
On a similar theme, digital leaders told us that some of the broader enablers of AI implementation aren’t in the right place either. When it comes to questions of funding, there’s too much hype around the potential benefits (whether cashable or not) that AI will offer in the near-term. That’s partly driven by external work from think tanks and others, with those in government offering a much more cautionary voice.
But interviewees told us that there needs to be a better understanding that AI is more than just a one-off investment – training, refining and improving algorithms will require cycles of funding. Recognising this will make it easier to properly evaluate the likely benefits AI could eventually provide.
Even more important is getting serious about data interoperability. This has long been a perennial Whitehall problem, but it’s even more important now. An AI assistant to help citizens use government services is less useful if it can access your health data and services – but can’t use any data held by every other government department.
Whitehall’s AI training offer needs work
Kevin Cunnington, the co-author of the report and former director general of the Government Digital Service (GDS), emphasised the role training played in supporting the UK government’s wave of digital transition. In fact, he notes in the Digital Leaders 2024 report: “GDS trained over 10,000 staff” to support that ambition. By contrast, digital leaders told us there was nothing like this kind of approach around AI yet.
There is some good work taking place. Some permanent secretaries attended an ‘away day’ at Oxford University to learn more about AI. One government department offers teach-ins where top officials hear from internal digital experts. In places like the government’s AI incubator (i.AI), valuable talent has been attracted and retained, offering real skills and expertise.
However, the overarching perception was that Whitehall has not gone far enough. Talking only about top officials, one interviewee said that they “know AI’s a big bet” but “there’s no materiality” beyond that. Another added that Whitehall leaders are around five years behind private sector CEOs: “There’s a big digital gap and then [there’s] an even bigger AI understanding gap.”
On the workforce more generally, we were told that the available training offer is essentially “unusable”. There’s nothing close to that 10,000 people ambition we saw around digital. And, though it’s now under development, the Government Digital and Data Profession Capability Framework – which defines tech roles and the skills that underpin them – hadn’t yet been updated to reflect AI competencies either.
Read more: UK Cabinet Office chief flags missions on civil service pay, digital tools and training
There are some reasons to be cheerful
This may seem pretty doom and gloom. And, in reality, our report presents a somewhat pessimistic picture. We’re much further behind than we could be. Many of the fundamentals – the core enablers to implementing AI – aren’t in the right place.
But there are some reasons to be cheerful. First, one of the key issues interviewees raised – the lack of a single AI voice in government – was addressed after the recent election. The new government announced that it was consolidating the CDDO, i.AI and GDS within the Department for Science, Innovation and Technology (DSIT). It’s clear that this will be welcome news to digital leaders across government.
Secondly, the UK has significant untapped potential. Oxford Insights ranks the UK third in its International AI Readiness Index. Tortoise Media places us fourth in its own Global AI Index. There is a central AI vision already developed (awaiting political sign-off), and the former Chair of the Advanced Research and Invention Agency Matt Clifford is developing a new AI Opportunities Action Plan. These are all positive signs.
And finally – to borrow a point made by Kevin Cunnington – we’ve been here before. Our report concludes by offering a roadmap for the UK’s AI transition, drawing on lessons from previous transformation programmes that accelerated digital capabilities in government.
The UK has made enormous progress around the use of digital technology in government in recent years. With AI – given all its potential and all the excitement around it – there’s no doubt that we have the ability to do the same, if we take the right actions. We just need to hit the accelerator, rather than focusing mostly on the brake.
About James Sweetland
James Sweetland is a researcher, journalist and consultant who works on government reform, policing and tech policy. He is the co-author of the Digital Leaders 2024 study.
Related Posts
